Proteins in granules

Cambridge Centre for Proteomics

Abstract

Here, we explore the potential granule-localised proteins

library(MSnbase)

Loading required package: BiocGenerics
Loading required package: parallel

Attaching package: ‘BiocGenerics’

The following objects are masked from ‘package:parallel’:

    clusterApply, clusterApplyLB, clusterCall, clusterEvalQ, clusterExport, clusterMap, parApply,
    parCapply, parLapply, parLapplyLB, parRapply, parSapply, parSapplyLB

The following objects are masked from ‘package:stats’:

    IQR, mad, sd, var, xtabs

The following objects are masked from ‘package:base’:

    anyDuplicated, append, as.data.frame, basename, cbind, colnames, dirname, do.call, duplicated,
    eval, evalq, Filter, Find, get, grep, grepl, intersect, is.unsorted, lapply, Map, mapply, match,
    mget, order, paste, pmax, pmax.int, pmin, pmin.int, Position, rank, rbind, Reduce, rownames,
    sapply, setdiff, sort, table, tapply, union, unique, unsplit, which, which.max, which.min

Loading required package: Biobase
Welcome to Bioconductor

    Vignettes contain introductory material; view with 'browseVignettes()'. To cite Bioconductor,
    see 'citation("Biobase")', and for packages 'citation("pkgname")'.

Loading required package: mzR
Loading required package: Rcpp
Loading required package: S4Vectors
Loading required package: stats4

Attaching package: ‘S4Vectors’

The following object is masked from ‘package:base’:

    expand.grid

Loading required package: ProtGenerics

Attaching package: ‘ProtGenerics’

The following object is masked from ‘package:stats’:

    smooth


This is MSnbase version 2.14.2 
  Visit https://lgatto.github.io/MSnbase/ to get started.


Attaching package: ‘MSnbase’

The following object is masked from ‘package:base’:

    trimws

library(pRoloc)

Loading required package: MLInterfaces
Loading required package: annotate
Loading required package: AnnotationDbi
Loading required package: IRanges
Loading required package: XML

Attaching package: ‘annotate’

The following object is masked from ‘package:mzR’:

    nChrom

Loading required package: cluster
Loading required package: BiocParallel
Registered S3 methods overwritten by 'dbplyr':
  method         from
  print.tbl_lazy     
  print.tbl_sql      
Registered S3 method overwritten by 'data.table':
  method           from
  print.data.table     

This is pRoloc version 1.28.0 
  Visit https://lgatto.github.io/pRoloc/ to get started.

library(pRolocExt)
library(tidyverse)

── Attaching packages ──────────────────────────────────────────────────────────────────────── tidyverse 1.3.0 ──
✓ ggplot2 3.3.2     ✓ purrr   0.3.4
✓ tibble  3.0.3     ✓ dplyr   1.0.4
✓ tidyr   1.1.2     ✓ stringr 1.4.0
✓ readr   1.3.1     ✓ forcats 0.5.0
── Conflicts ─────────────────────────────────────────────────────────────────────────── tidyverse_conflicts() ──
x dplyr::collapse()   masks IRanges::collapse()
x dplyr::combine()    masks MSnbase::combine(), Biobase::combine(), BiocGenerics::combine()
x dplyr::desc()       masks IRanges::desc()
x tidyr::expand()     masks S4Vectors::expand()
x dplyr::filter()     masks stats::filter()
x dplyr::first()      masks S4Vectors::first()
x dplyr::lag()        masks stats::lag()
x ggplot2::Position() masks BiocGenerics::Position(), base::Position()
x purrr::reduce()     masks IRanges::reduce(), MSnbase::reduce()
x dplyr::rename()     masks S4Vectors::rename()
x dplyr::select()     masks AnnotationDbi::select()
x dplyr::slice()      masks IRanges::slice()

library(camprotR)
library(biobroom)

Loading required package: broom
Registered S3 methods overwritten by 'biobroom':
  method      from 
  glance.list broom
  tidy.list   broom

source('../plot_foi.R')

colours <- readRDS('../../../../6_shiny_app/out/shiny_colours.rds')$Protein

Read in BANDLE results and fully annotated LOPIT data

diff_loc <- readRDS('../../out/bandle_diff_loc_all_unique.rds')

combined_protein_res_inc_bandle <- readRDS('../../out/combined_protein_res_inc_bandle_loc.rds')

Just looking at the profile for ZNF622

marker_profiles <- combined_protein_res_inc_bandle %>% names() %>%
  lapply(function(name){
    mrkConsProfiles(combined_protein_res_inc_bandle[[name]])[c('ER', 'RIBOSOME'),] %>%
      data.frame() %>%
      tibble::rownames_to_column('markers') %>%
      pivot_longer(cols=-markers, names_to='sample') %>%
      mutate(sample=remove_x(sample)) %>%
      mutate(condition=name) %>%
      merge(pData(combined_protein_res_inc_bandle[[name]])[,c('fraction', 'replicate')],
            by.x='sample', by.y='row.names')
  }) %>% bind_rows() %>%
  mutate(markers=update_loc_names(markers))

Registered S3 method overwritten by 'htmlwidgets':
  method           from         
  print.htmlwidget tools:rstudio

foi_profile <- combined_protein_res_inc_bandle %>%
  lapply(function(x) tidy(x['Q969S3',], addPheno=TRUE)) %>%
  bind_rows() %>%
  mutate(markers='ZNF622')

`tbl_df()` is deprecated as of dplyr 1.0.0.
Please use `tibble::as_tibble()` instead.
This warning is displayed once every 8 hours.
Call `lifecycle::last_warnings()` to see where this warning was generated.

p <- bind_rows(foi_profile, marker_profiles) %>%
  mutate(condition=recode(condition, 'Thapsigargin'='UPR')) %>%
  ggplot(aes(fraction, value, group=interaction(replicate, markers), colour=markers)) +
  geom_line(size=1) +
  facet_grid(condition~(paste0('Replicate ', replicate))) +
  scale_x_continuous(breaks=1:8, name='Fraction') +
  theme_camprot(base_size=15, base_family='sans', border=FALSE) +
  scale_colour_manual(values=c(colours[c('ER', 'RIBOSOME')], get_cat_palette(7)[7]), name='') +
  theme(strip.background=element_blank()) +
  xlab('Fraction') +
  ylab('Abundance (sum norm.)')

print(p)
ggsave('../../../../5_manuscript_figures/Figure_5/model/znf622.png', width=7, height=3.5)

ggsave('../../../../5_manuscript_figures/Figure_5/model/znf622.pdf', width=7, height=3.5)

NA
NA
NA
NA

ZNF622 and CAMK2D

foi_for_validation <- c('Q969S3', 'Q13557')
compare_profiles(foi_for_validation)

diff_loc %>% filter(protein %in% foi_for_validation)

Now, onto the proteins moving away from the ribosome.

bandle_from_ribosome <- diff_loc %>%
  filter(bandle.allocation.dmso.minimal=='RIBOSOME', level!='Candidate')

print(bandle_from_ribosome)

compare_profiles(bandle_from_ribosome$protein)

plot_fois(as.character(bandle_from_ribosome$protein), foi_name='Away from ribosome', feature_col='bandle_alloc',
          moi=c('Cytosol', 'Ribosome', 'ER', 'Nucleus', 'Protein Complex'), plot_tsne=TRUE, unknown_desc='Undefined')

NA
NA

PABPC1, PABPC4, UPF1 & STAU2 seem quite likely to be moving to a Stress Granule (SG) profile given a) They are associated with ribosomes when translation is active, b) most are known to be associated with SG upon stress & c) they increase in abundance in the fractions indicative of granules in the LoRNA data (fraction 5 & 6).

Whereas, MAGEB2 & PA2G4 seem to move towards cytosol if anything.

PA2G4 - regulates cap-independent IRES-mediated translation

Can we identify further proteins which are similarly localised in Tg by just using correlation with the above.

potential_sg_markers <- bandle_from_ribosome %>% filter(grepl('(PABP|UPF1|STAU2)', `GENES`)) %>% pull(protein)

p <- compare_profiles(potential_sg_markers)
print(p)

ggsave('../../../../5_manuscript_figures/Figure_4/granules/bandle_reloc.png', width=5, height=5)

ggsave('../../../../5_manuscript_figures/Figure_4/granules/bandle_reloc.pdf', width=5, height=5)

plot_fois(potential_sg_markers, foi_name='PABPC1/4, UPF1, STAU2', feature_col='bandle_alloc',
          moi=c('Cytosol', 'Nucleus', 'Ribosome'), plot_tsne=TRUE, unknown_desc='Undefined', )

NA

plot_tsne(combined_protein_res_inc_bandle$DMSO, 'markers')

plot_tsne(combined_protein_res_inc_bandle$Thapsigargin, 'markers')

plot_tsne(combined_protein_res_inc_bandle$DMSO, 'bandle_alloc', unknown='Undefined')

plot_tsne(combined_protein_res_inc_bandle$Thapsigargin, 'bandle_alloc', unknown='Undefined')

plot_tsne(combined_protein_res_inc_bandle$DMSO, 'markers', foi=potential_sg_markers)

plot_tsne(combined_protein_res_inc_bandle$Thapsigargin, 'markers', foi=potential_sg_markers)

plot_tsne(combined_protein_res_inc_bandle$DMSO, 'bandle_alloc', foi=potential_sg_markers, unknown='Undefined')

plot_tsne(combined_protein_res_inc_bandle$Thapsigargin, 'bandle_alloc', foi=potential_sg_markers, unknown='Undefined')

Calculate all vs all correlations and then identify potential new SG proteins as those top 1% most highly correlated with our 4 proteins, plus correlation > 0.35 greater than correlation with ribosomes.

all_cor <- combined_protein_res_inc_bandle[[2]] %>% exprs() %>% t() %>% cor(method='pearson')

wh <- which(rownames(all_cor) %in% potential_sg_markers)

mean_cor_with_potential_sg <- all_cor[,wh] %>% rowMeans()



mean_cor_with_potential_sg[wh]

   P11940    Q13310    Q92900    Q9NUL3 
0.9322607 0.9185716 0.8253327 0.8607352 

mean_cor_with_ribosome <- all_cor[,which(getMarkers(combined_protein_res_inc_bandle[[2]])=='RIBOSOME')] %>% rowMeans()

organelleMarkers
        CYTOSOL              ER           GOLGI        LYSOSOME    MITOCHONDRIA   NUCLEOPLASM-1   NUCLEOPLASM-2 
             68              83              35              27             150              27              19 
        NUCLEUS      PEROXISOME              PM PROTEIN COMPLEX        RIBOSOME         unknown 
             74              14              57              62              66            4632 

hist(mean_cor_with_potential_sg)

hist(mean_cor_with_potential_sg-mean_cor_with_ribosome)

summary(mean_cor_with_potential_sg-mean_cor_with_ribosome)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
-0.52672 -0.38363 -0.00206 -0.02010  0.25158  0.68978 

potential_new_sg <- names(mean_cor_with_potential_sg)[
  ((mean_cor_with_potential_sg-mean_cor_with_ribosome)>0.35 &
     mean_cor_with_potential_sg > quantile(mean_cor_with_potential_sg, 0.99))] %>% 
  setdiff(potential_sg_markers)

Looking at localisation of thse potential SG proteins.

plot_fois(potential_new_sg, foi_name='Potential SG',
          moi=c('CYTOSOL', 'RIBOSOME', 'ER', 'NUCLEUS', 'PROTEIN COMPLEX'),
          feature_col='bandle_alloc',
          plot_tsne=TRUE, unknown_desc='Undefined')

NA
NA

What are the 22 potential new SG proteins…

fData(combined_protein_res_inc_bandle$Thapsigargin)[potential_new_sg,174:182]

Profiles for all 22 proteins

compare_profiles(potential_new_sg) + facet_wrap(~name, ncol=6)

OK, let’s note down what we think about the above 22 proteins

Confident RNA granule proteins (PB and/or SG) = 10 proteins

CASC3 (https://www.uniprot.org/uniprot/O15234) - Recruited to SG (https://pubmed.ncbi.nlm.nih.gov/17652158/)

CNOT1,2,3,6L,9 - Part of CCR4-NOT complex which is known to be localised to RNP granules (e.g https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3376659/)

DCP1A, B (https://www.uniprot.org/uniprot/Q9NPI6; https://www.uniprot.org/uniprot/Q8IZD4) - Necessary for the degradation of mRNAs, both in normal mRNA turnover and in nonsense-mediated mRNA decay. Removes the 7-methyl guanine cap structure from mRNA molecules, yielding a 5’-phosphorylated mRNA fragment and 7m-GDP. Contributes to the transactivation of target genes after stimulation by TGFB1. PB localised

SECISBP2 (https://www.uniprot.org/uniprot/Q96T21) - Binds to the SECIS element in the 3’-UTR of some mRNAs encoding selenoproteins. Binding is stimulated by SELB. Found in SG (https://en.wikipedia.org/wiki/Stress_granule, inc Youn et al)

YBX3 https://www.uniprot.org/uniprot/P16989 YBX1 - Promotes mRNA stabilization: acts by binding to m5C-containing mRNAs and recruiting the mRNA stability maintainer ELAVL1, thereby preventing mRNA decay. Both YBX1 & 3 are often found in SG (https://en.wikipedia.org/wiki/Stress_granule, inc Youn et al for YBX3).

Plausible though undetected = 2 proteins

CAMK2D (https://www.uniprot.org/uniprot/Q13557) - Calcium/calmodulin-dependent protein kinase involved in the regulation of Ca2+ homeostatis!! Seems to be sarcoplasmic reticulum more than ER localised. CAMK2 has been implicated in RNP decondensation (https://elifesciences.org/articles/65742) so this is at least plausible granule localised

RANB9 (https://www.uniprot.org/uniprot/Q96S59) - Inhibits FMR1 binding to RNA (https://pubmed.ncbi.nlm.nih.gov/15381419/) and other RAN binding proteins are found in SG (https://en.wikipedia.org/wiki/Stress_granule) so plausible if not already known

CAMK2D & RANB9 are worth reading into more!

Implausible = 10 proteins

GIT1/2 (https://www.uniprot.org/uniprot/Q9Y2X7; https://www.uniprot.org/uniprot/Q14161) - Seems to localised in focal adhesions

HAUS3 (https://www.uniprot.org/uniprot/Q68CZ6) - Contributes to mitotic spindle assembly, maintenance of centrosome integrity and completion of cytokinesis as part of the HAUS augmin-like complex.

MKLN1 (https://www.uniprot.org/uniprot/Q9UL63) - Component of the CTLH E3 ubiquitin-protein ligase complex that selectively accepts ubiquitin from UBE2H and mediates ubiquitination and subsequent proteasomal degradation of the transcription factor HBP1

PRPS2 & PRPSAP1 (https://www.uniprot.org/uniprot/P11908; https://www.uniprot.org/uniprot/Q14558) - Catalyzes the synthesis of phosphoribosylpyrophosphate (PRPP) that is essential for nucleotide synthesis; Seems to play a negative regulatory role in 5-phosphoribose 1-diphosphate synthesis.

RMND5A (https://www.uniprot.org/uniprot/Q9H871) - Core component of the CTLH E3 ubiquitin-protein ligase

RNF213 (https://www.uniprot.org/uniprot/Q63HN8) - E3 ubiquitin-protein ligase involved in angiogenesis. Other RNFs (214, 219, 25) have been detected in SGs before (https://en.wikipedia.org/wiki/Stress_granule)

SEPTIN6/8 (https://www.uniprot.org/uniprot/Q14141; https://www.uniprot.org/uniprot/Q92599) - Filament-forming cytoskeletal GTPases. Required for normal organization of the actin cytoskeleton.

Explaining the above:

• I can’t find any convincing link between FA and PB/SG

• Lots of ubiquitin-related proteins have been observed in SG before, including ubiquitin-protein ligasess (https://en.wikipedia.org/wiki/Stress_granule; DTL, DTX3L, TRIM21, TRIM25, TRIM56, TRIM71, UBB & UBL5). However, ubiquitination is dispensable to SG formation and dissolution: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6508666/

My interpretation: RMND5A may be worth following up but the others look likely to be FPs.

10/22 proteins being strongly P-body/SG does suggest we may be resolving macromolecular content of the granule in our density gradient!